As Prepared for Delivery
Good afternoon and thanks for having me here today. The topic of today’s forum – the resilience of Treasury markets – is a topic that we at the Treasury Department care deeply about and accordingly spend a great deal of time focused on. And the title of today’s forum – “The Path to Resilient Treasury Markets” – seems exactly right to me. The use of the word “path” reminds me of the efforts undertaken over most of the past decade. Resilient Treasury markets are not a static endpoint; they are a journey towards a shared goal and responsibility of the Treasury, the broader official sector, and market participants.
Over the last few years, Treasury and our colleagues in the Inter-Agency Working Group on Treasury Market Surveillance, or IAWG, have made significant progress in developing policies that would help make the Treasury market more resilient, but there is still much work to be done. And as the structure of the market evolves, our policies will need to evolve as well. Venues like today’s forum are important opportunities for the official sector and market participants to discuss progress and share insights on the best paths forward.
While IAWG members continue to consider a wide range of policies to help make the Treasury markets more resilient, my remarks today will focus on one: buybacks. After careful consideration, we decided it was prudent to move forward and announced our intentions at the May refunding to implement a regular buyback program next year.1 We believe buybacks can play an important role in helping to make the Treasury market more liquid and resilient by providing liquidity support. The buyback program will also help Treasury to better achieve our debt management objectives. Before delving into our upcoming plans, I will first briefly discuss Treasury’s recent history with buybacks.
The last time Treasury conducted a regular buyback program was in the early 2000s, ending in April 2002. To understand why these buybacks were conducted, it’s important to review what was happening with Treasury auction sizes leading up to those purchases. Due to shrinking budget deficits and eventual surpluses in the 1990s, Treasury significantly reduced offering sizes of the 2- and 3-year notes between 1996 and 1998. As the outstanding sizes of these issues continued to shrink, on-the-run liquidity was hindered. This led to concerns of lower auction demand and therefore higher financing costs for Treasury. Market observers at the time cautioned that Treasury securities were at risk of losing “benchmark status” because the auction sizes were becoming too small. To delay shrinking new issue sizes further, in 1998, Treasury discontinued the 3-year offering and changed the 5-year offering from monthly to quarterly. Another consequence of these smaller short- and intermediate-term offerings, however, was an unintended increase in the average maturity of debt outstanding, as long-term debt remained outstanding. At the time, this was considered potentially costly for Treasury, based on the assumption that the yield curve is typically positively sloped.
Since budget surpluses were expected to persist into the foreseeable future, at the end of the 1990s Treasury began exploring options to address concerns regarding shrinking auction sizes, including using buybacks to maintain larger new issue sizes. After careful consideration of a range of options, including recommendations from the Treasury Borrowing Advisory Committee, or TBAC, Treasury adopted a final rule on buybacks in January 2000. At the time of adoption, then-Treasury Secretary Summers noted several advantages of buybacks, including (1) “to enhance liquidity of benchmark securities”, (2) “to prevent what would otherwise be a potentially costly and unjustified increase in the average maturity of our debt”, and (3) “more effective use of excess cash”.2 Between March 2000 and April 2002, Treasury conducted 45 buyback operations, buying back a total of $67.5 billion of outstanding debt. The first 42 operations were intended to address concerns related to shrinking auction sizes. However, as budget surpluses disappeared, in April 2002 the last three buyback operations were aimed at lowering high seasonal cash balances.
While Treasury has conducted limited small-value buybacks each year for most of the past decade to test operational capabilities and has always considered them as part of our debt management toolkit, it has been a little over 20 years since we last conducted buybacks in meaningful size. So why the change? Why reintroduce buybacks now?
First, a buyback program in the Treasury market would not be unique. Many sovereign debt management offices around the world use them to achieve various objectives. A 2022 OECD survey of nearly 40 sovereign DMOs indicated that a majority of the respondents conduct buybacks or switches (that is, exchanges of one security for another).3 Additionally, there have been a handful of studies suggesting there are both liquidity and cash management benefits of conducting buybacks, even when there are budget deficits.4 Finally, we have received a significant amount of feedback over the years, including from the TBAC, the primary dealers, and other market participants, supportive of a regular buyback program.
We have two objectives for our upcoming buyback program: liquidity support and cash management. Before expanding on these, let me be clear on what this program is not intended for. First, we do not intend to conduct tactical or ad-hoc operations; instead, buybacks will be conducted in a regular and predictable manner. Second, we do not intend to use buybacks to change the maturity profile of our debt outstanding. Our issuance strategy remains our primary tool for adjusting the maturity profile, if desired, and we plan to limit the size of our buyback operations and conduct operations across the curve. And third, buybacks are not intended to address acute periods of market stress – I’ll come back to this in a moment.
As I noted, the first objective of our buyback program is liquidity support. We believe buybacks can help improve the liquidity of the Treasury market by providing a regular opportunity for market participants to sell back to Treasury off-the-run securities across the yield curve. This should improve the willingness of investors and intermediaries to trade and provide liquidity in these securities, all else equal, knowing there is a potential outlet to sell some of their off-the-run holdings. These operations can also make intermediation capacity more readily available by Treasury purchasing hard-to-move securities that are in broker-dealer inventories and use up their intermediation capacity. With more intermediation capacity readily available, the Treasury market should see improved functioning in normal times and have an enhanced ability to absorb larger flows.
While we believe it is important that we retain some flexibility in providing incremental liquidity support to certain sectors of the Treasury market, it is important we make it clear at the outset that these buybacks are not intended to ameliorate periods of acute market stress. Unlike the Federal Reserve System, which can finance purchases of securities by creating reserves, each dollar of buybacks needs to be financed with a dollar of Treasury issuance, all else equal. This limits our ability to rapidly increase the size of buybacks to a level potentially necessary to alleviate market stress without resulting in significant costs for the taxpayer, as a corresponding rapid rise in Treasury issuance could materially increase our financing costs during these periods.
The second objective of our buyback program is cash management. We believe buybacks can improve our cash management by reducing the volatility in our cash balance and bill issuance in two ways. First, during periods of high tax receipts, our cash balance typically rises meaningfully, and we try to offset some of this rise by making significant reductions in bill issuance. This volatility can be costly to the taxpayer by causing imbalances between supply and demand and by potentially hindering the smooth functioning of the bill market. Conducting buybacks during these periods could help reduce some of this volatility. Second, we plan to focus cash management buybacks on purchasing securities with maturity dates around periods of large outflows, which could also reduce this volatility. These buybacks could mitigate the need to raise extra cash through bill issuance to prepare for such periods.
Operationally, both liquidity support and cash management buybacks will be executed by our fiscal agent, the Federal Reserve Bank of New York, via the same trading platform that they have used for the past 17 years to conduct their purchases of Treasury securities. While the goals of our buyback program are quite different from the goals of the Federal Reserve System’s prior Treasury securities purchase programs, the actual operations should be familiar to market participants because many of the design parameters are similar. However, an important difference is that we intend to be more price sensitive in selecting the buyback offers to accept. Because we are not targeting a fixed quantity to purchase, we believe it is important to only buy back securities that help us meet our objectives and to only execute at attractive prices. As a result, the amount we buy during any given operation might be materially lower than the maximum amount announced. The amount that we buy will be closely tied to the quality of the offers we receive.
To measure the success of the buyback program, we plan to utilize both quantitative and qualitative measures; there is no single measure that we plan to rely on. For liquidity support buybacks, while we hope to see some improvement across liquidity metrics over time, we recognize there are many other factors that may affect liquidity at any given moment, such as market volatility. It has also been a while since we last conducted regular buyback operations; as such, our understanding of the effects of buybacks will improve as we conduct more operations and have additional data for analysis. We are confident that the operational design parameters we have communicated thus far are a good starting point, but over time we may need to make some adjustments to better achieve our objectives.5
Turning to one last question on buybacks: how should they be financed? We plan to treat the increase in our borrowing needs due to buybacks the same way that we treat other outlays. Moreover, we think it would be imprudent to try to replace the interest rate risk that buybacks remove from a certain maturity sector in the market by precisely offsetting them with issuance of an equivalent amount of on-the-run Treasury securities in the same maturity sector. Attempts to do so would unnecessarily limit the flexibility of our issuance strategy and run counter to regular and predictable issuance. When evaluating how buybacks affect our issuance plans, we will continue to look at our borrowing needs in totality in the short-, intermediate-, and long-term and make issuance adjustments based on what we believe is the appropriate mix of coupon and bill issuance. Unexpected short-term changes in borrowing needs will still largely be addressed with changes in bill issuance.
Before I conclude, I wanted to close with a few words about recent debt management decisions. Since the suspension of the debt limit in early June, we have substantially increased gross issuance, both to rebuild the Treasury General Account from very low levels and to continue to meet the government’s ongoing borrowing needs. Nearly all of this increase has been in Treasury bills, where more than $1 trillion of increased supply has been absorbed well. The total supply of bills now exceeds $5.2 trillion, which represents just over 20 percent of marketable Treasury debt outstanding.
While bill supply is likely to remain above 20 percent for some time, our intention is to gradually increase coupon issuance to better align auction sizes with intermediate- to long-term borrowing needs, and as I outlined at the August refunding, we expect that further gradual increases in coupon auctions sizes will likely be necessary in future quarters. The TBAC has supported this approach, recommending at its August meeting that considering the scale and uncertainty of projected borrowing needs, the amount of bills as a share of total marketable debt outstanding should deviate from TBAC’s recommended 15-20 percent range for some time.
Treasury will continue to follow its longstanding “regular and predictable” approach to issuance and continue to use our quarterly refunding statements to provide updates on future changes to borrowing needs and issuance. Our approach will continue to depend on a variety of factors, including the evolution of the fiscal outlook and the pace and duration of future redemptions from the Federal Reserve’s SOMA portfolio.
In conclusion, I am hopeful that the upcoming regular buyback program will become an important feature of the Treasury market, improving liquidity and helping Treasury meet our debt management objectives. The feedback we have received to date from a wide variety of market participants has been tremendously helpful and will continue to be important as we further refine the operational design parameters of the program ahead of and after our planned 2024 launch. Enhancing the resilience of the Treasury market remains a top priority of the official sector, and I look forward to the rest of today’s discussion on ways to ensure that the Treasury market remains the deepest and most liquid market in the world. Thank you.
###
1. https://home.treasury.gov/news/press-releases/jy1460
2. https://home.treasury.gov/news/press-releases/ls330
3. https://www.oecd-ilibrary.org/sites/62c91ab0-en/index.html?itemId=/content/component/62c91ab0-en
4. https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr304.pdf and https://www.gao.gov/assets/gao-12-314.pdf
5. For a more detailed discussion of design issues, see https://home.treasury.gov/system/files/221/TreasurySupplementalQRQ22023.pdf and https://home.treasury.gov/system/files/221/TreasurySupplementalQRQ32023.pdf